Hugging Face's logo

CharlesLi
/
llama_2_alpaca_cot_simplest

PEFT
TensorBoard
Safetensors
llama
alignment-handbook
trl
sft
Generated from Trainer
Model card Files
xet
 Metrics Community
CharlesLi commited on
Commit
0df5e96
·
verified ·
1 Parent(s): 6aa7207

Model save

Browse files
Files changed (4) hide show
  1. README.md +11 -12
  2. all_results.json +6 -11
  3. train_results.json +6 -6
  4. trainer_state.json +85 -85
README.md CHANGED
@@ -3,7 +3,6 @@ base_model: meta-llama/Llama-2-7b-chat-hf
3
  library_name: peft
4
  license: llama2
5
  tags:
6
- - alignment-handbook
7
  - trl
8
  - sft
9
  - generated_from_trainer
@@ -19,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
19
 
20
  This model is a fine-tuned version of [meta-llama/Llama-2-7b-chat-hf](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) on the None dataset.
21
  It achieves the following results on the evaluation set:
22
- - Loss: 0.7487
23
 
24
  ## Model description
25
 
@@ -56,16 +55,16 @@ The following hyperparameters were used during training:
56
 
57
  | Training Loss | Epoch | Step | Validation Loss |
58
  |:-------------:|:------:|:----:|:---------------:|
59
- | 1.9905 | 0.7692 | 5 | 1.6990 |
60
- | 1.1835 | 1.5385 | 10 | 1.1457 |
61
- | 1.0343 | 2.3077 | 15 | 1.0081 |
62
- | 1.0226 | 3.0769 | 20 | 0.9697 |
63
- | 0.9336 | 4.0769 | 25 | 0.8462 |
64
- | 0.8757 | 4.8462 | 30 | 0.8025 |
65
- | 0.8223 | 5.6154 | 35 | 0.7699 |
66
- | 0.791 | 6.3846 | 40 | 0.7559 |
67
- | 0.7922 | 7.1538 | 45 | 0.7494 |
68
- | 0.7706 | 7.9231 | 50 | 0.7487 |
69
 
70
 
71
  ### Framework versions
 
3
  library_name: peft
4
  license: llama2
5
  tags:
 
6
  - trl
7
  - sft
8
  - generated_from_trainer
 
18
 
19
  This model is a fine-tuned version of [meta-llama/Llama-2-7b-chat-hf](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) on the None dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 0.7382
22
 
23
  ## Model description
24
 
 
55
 
56
  | Training Loss | Epoch | Step | Validation Loss |
57
  |:-------------:|:------:|:----:|:---------------:|
58
+ | 2.1392 | 0.7692 | 5 | 2.1469 |
59
+ | 1.2925 | 1.5385 | 10 | 1.1819 |
60
+ | 1.0332 | 2.3077 | 15 | 0.8971 |
61
+ | 0.9534 | 3.0769 | 20 | 0.8235 |
62
+ | 0.8612 | 3.8462 | 25 | 0.7759 |
63
+ | 0.7801 | 4.6154 | 30 | 0.7457 |
64
+ | 0.7752 | 5.3846 | 35 | 0.7379 |
65
+ | 0.6727 | 6.1538 | 40 | 0.7387 |
66
+ | 0.7328 | 6.9231 | 45 | 0.7381 |
67
+ | 0.6869 | 7.6923 | 50 | 0.7382 |
68
 
69
 
70
  ### Framework versions
all_results.json CHANGED
@@ -1,14 +1,9 @@
1
  {
2
- "epoch": 7.923076923076923,
3
- "eval_loss": 0.7486859560012817,
4
- "eval_runtime": 1.3731,
5
- "eval_samples": 20,
6
- "eval_samples_per_second": 14.565,
7
- "eval_steps_per_second": 2.185,
8
- "total_flos": 1436866805760.0,
9
- "train_loss": 0.4985464477539063,
10
- "train_runtime": 119.1701,
11
  "train_samples": 100,
12
- "train_samples_per_second": 6.713,
13
- "train_steps_per_second": 0.42
14
  }
 
1
  {
2
+ "epoch": 7.6923076923076925,
3
+ "total_flos": 1447302955008.0,
4
+ "train_loss": 0.982821364402771,
5
+ "train_runtime": 185.3525,
 
 
 
 
 
6
  "train_samples": 100,
7
+ "train_samples_per_second": 4.316,
8
+ "train_steps_per_second": 0.27
9
  }
train_results.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
- "epoch": 7.923076923076923,
3
- "total_flos": 1436866805760.0,
4
- "train_loss": 0.4985464477539063,
5
- "train_runtime": 119.1701,
6
  "train_samples": 100,
7
- "train_samples_per_second": 6.713,
8
- "train_steps_per_second": 0.42
9
  }
 
1
  {
2
+ "epoch": 7.6923076923076925,
3
+ "total_flos": 1447302955008.0,
4
+ "train_loss": 0.982821364402771,
5
+ "train_runtime": 185.3525,
6
  "train_samples": 100,
7
+ "train_samples_per_second": 4.316,
8
+ "train_steps_per_second": 0.27
9
  }
trainer_state.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
- "epoch": 7.923076923076923,
5
  "eval_steps": 5,
6
  "global_step": 50,
7
  "is_hyper_param_search": false,
@@ -10,169 +10,169 @@
10
  "log_history": [
11
  {
12
  "epoch": 0.15384615384615385,
13
- "grad_norm": 0.798818668071787,
14
- "learning_rate": 0.0001,
15
  "loss": 1.6442,
16
  "step": 1
17
  },
18
  {
19
  "epoch": 0.7692307692307693,
20
- "grad_norm": 1.3386421604346492,
21
- "learning_rate": 0.00018660254037844388,
22
- "loss": 1.9905,
23
  "step": 5
24
  },
25
  {
26
  "epoch": 0.7692307692307693,
27
- "eval_loss": 1.698988676071167,
28
- "eval_runtime": 2.6255,
29
- "eval_samples_per_second": 7.617,
30
- "eval_steps_per_second": 1.143,
31
  "step": 5
32
  },
33
  {
34
  "epoch": 1.5384615384615383,
35
- "grad_norm": 0.7866298084520749,
36
- "learning_rate": 0.00011736481776669306,
37
- "loss": 1.1835,
38
  "step": 10
39
  },
40
  {
41
  "epoch": 1.5384615384615383,
42
- "eval_loss": 1.1456953287124634,
43
- "eval_runtime": 1.3505,
44
- "eval_samples_per_second": 14.809,
45
- "eval_steps_per_second": 2.221,
46
  "step": 10
47
  },
48
  {
49
  "epoch": 2.3076923076923075,
50
- "grad_norm": 0.48543322396053706,
51
- "learning_rate": 3.5721239031346066e-05,
52
- "loss": 1.0343,
53
  "step": 15
54
  },
55
  {
56
  "epoch": 2.3076923076923075,
57
- "eval_loss": 1.008058786392212,
58
- "eval_runtime": 1.34,
59
- "eval_samples_per_second": 14.925,
60
  "eval_steps_per_second": 2.239,
61
  "step": 15
62
  },
63
  {
64
  "epoch": 3.076923076923077,
65
- "grad_norm": 0.5909954874779454,
66
- "learning_rate": 0.0,
67
- "loss": 1.0226,
68
  "step": 20
69
  },
70
  {
71
  "epoch": 3.076923076923077,
72
- "eval_loss": 0.969663143157959,
73
- "eval_runtime": 1.3374,
74
- "eval_samples_per_second": 14.955,
75
- "eval_steps_per_second": 2.243,
76
  "step": 20
77
  },
78
  {
79
- "epoch": 4.076923076923077,
80
- "grad_norm": 0.8718791279173701,
81
  "learning_rate": 0.00011736481776669306,
82
- "loss": 0.9336,
83
  "step": 25
84
  },
85
  {
86
- "epoch": 4.076923076923077,
87
- "eval_loss": 0.8462212681770325,
88
- "eval_runtime": 2.6035,
89
- "eval_samples_per_second": 7.682,
90
- "eval_steps_per_second": 1.152,
91
  "step": 25
92
  },
93
  {
94
- "epoch": 4.846153846153846,
95
- "grad_norm": 0.9304701683409617,
96
  "learning_rate": 8.263518223330697e-05,
97
- "loss": 0.8757,
98
  "step": 30
99
  },
100
  {
101
- "epoch": 4.846153846153846,
102
- "eval_loss": 0.8024625778198242,
103
- "eval_runtime": 1.3272,
104
- "eval_samples_per_second": 15.069,
105
- "eval_steps_per_second": 2.26,
106
  "step": 30
107
  },
108
  {
109
- "epoch": 5.615384615384615,
110
- "grad_norm": 0.4936722714962673,
111
  "learning_rate": 5.000000000000002e-05,
112
- "loss": 0.8223,
113
  "step": 35
114
  },
115
  {
116
- "epoch": 5.615384615384615,
117
- "eval_loss": 0.7699161767959595,
118
- "eval_runtime": 1.3288,
119
- "eval_samples_per_second": 15.051,
120
- "eval_steps_per_second": 2.258,
121
  "step": 35
122
  },
123
  {
124
- "epoch": 6.384615384615385,
125
- "grad_norm": 0.25077460367526005,
126
  "learning_rate": 2.339555568810221e-05,
127
- "loss": 0.791,
128
  "step": 40
129
  },
130
  {
131
- "epoch": 6.384615384615385,
132
- "eval_loss": 0.7559089660644531,
133
- "eval_runtime": 1.3364,
134
- "eval_samples_per_second": 14.966,
135
- "eval_steps_per_second": 2.245,
136
  "step": 40
137
  },
138
  {
139
- "epoch": 7.153846153846154,
140
- "grad_norm": 0.2752033434979242,
141
  "learning_rate": 6.030737921409169e-06,
142
- "loss": 0.7922,
143
  "step": 45
144
  },
145
  {
146
- "epoch": 7.153846153846154,
147
- "eval_loss": 0.7494038343429565,
148
- "eval_runtime": 1.3276,
149
- "eval_samples_per_second": 15.065,
150
- "eval_steps_per_second": 2.26,
151
  "step": 45
152
  },
153
  {
154
- "epoch": 7.923076923076923,
155
- "grad_norm": 0.34702034077494104,
156
  "learning_rate": 0.0,
157
- "loss": 0.7706,
158
  "step": 50
159
  },
160
  {
161
- "epoch": 7.923076923076923,
162
- "eval_loss": 0.7486859560012817,
163
- "eval_runtime": 1.3269,
164
- "eval_samples_per_second": 15.073,
165
- "eval_steps_per_second": 2.261,
166
  "step": 50
167
  },
168
  {
169
- "epoch": 7.923076923076923,
170
  "step": 50,
171
- "total_flos": 1436866805760.0,
172
- "train_loss": 0.4985464477539063,
173
- "train_runtime": 119.1701,
174
- "train_samples_per_second": 6.713,
175
- "train_steps_per_second": 0.42
176
  }
177
  ],
178
  "logging_steps": 5,
@@ -192,7 +192,7 @@
192
  "attributes": {}
193
  }
194
  },
195
- "total_flos": 1436866805760.0,
196
  "train_batch_size": 4,
197
  "trial_name": null,
198
  "trial_params": null
 
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
+ "epoch": 7.6923076923076925,
5
  "eval_steps": 5,
6
  "global_step": 50,
7
  "is_hyper_param_search": false,
 
10
  "log_history": [
11
  {
12
  "epoch": 0.15384615384615385,
13
+ "grad_norm": 0.7987578036789214,
14
+ "learning_rate": 4e-05,
15
  "loss": 1.6442,
16
  "step": 1
17
  },
18
  {
19
  "epoch": 0.7692307692307693,
20
+ "grad_norm": 1.4536572670447663,
21
+ "learning_rate": 0.0002,
22
+ "loss": 2.1392,
23
  "step": 5
24
  },
25
  {
26
  "epoch": 0.7692307692307693,
27
+ "eval_loss": 2.1469364166259766,
28
+ "eval_runtime": 2.606,
29
+ "eval_samples_per_second": 7.675,
30
+ "eval_steps_per_second": 1.151,
31
  "step": 5
32
  },
33
  {
34
  "epoch": 1.5384615384615383,
35
+ "grad_norm": 0.938799841460093,
36
+ "learning_rate": 0.00019396926207859084,
37
+ "loss": 1.2925,
38
  "step": 10
39
  },
40
  {
41
  "epoch": 1.5384615384615383,
42
+ "eval_loss": 1.181947112083435,
43
+ "eval_runtime": 1.3337,
44
+ "eval_samples_per_second": 14.996,
45
+ "eval_steps_per_second": 2.249,
46
  "step": 10
47
  },
48
  {
49
  "epoch": 2.3076923076923075,
50
+ "grad_norm": 0.5916866520464388,
51
+ "learning_rate": 0.0001766044443118978,
52
+ "loss": 1.0332,
53
  "step": 15
54
  },
55
  {
56
  "epoch": 2.3076923076923075,
57
+ "eval_loss": 0.8970538973808289,
58
+ "eval_runtime": 1.3399,
59
+ "eval_samples_per_second": 14.926,
60
  "eval_steps_per_second": 2.239,
61
  "step": 15
62
  },
63
  {
64
  "epoch": 3.076923076923077,
65
+ "grad_norm": 1.6899482115511137,
66
+ "learning_rate": 0.00015000000000000001,
67
+ "loss": 0.9534,
68
  "step": 20
69
  },
70
  {
71
  "epoch": 3.076923076923077,
72
+ "eval_loss": 0.8234956860542297,
73
+ "eval_runtime": 1.3359,
74
+ "eval_samples_per_second": 14.971,
75
+ "eval_steps_per_second": 2.246,
76
  "step": 20
77
  },
78
  {
79
+ "epoch": 3.8461538461538463,
80
+ "grad_norm": 0.6188790517019873,
81
  "learning_rate": 0.00011736481776669306,
82
+ "loss": 0.8612,
83
  "step": 25
84
  },
85
  {
86
+ "epoch": 3.8461538461538463,
87
+ "eval_loss": 0.7758801579475403,
88
+ "eval_runtime": 1.3323,
89
+ "eval_samples_per_second": 15.012,
90
+ "eval_steps_per_second": 2.252,
91
  "step": 25
92
  },
93
  {
94
+ "epoch": 4.615384615384615,
95
+ "grad_norm": 0.15897285891594412,
96
  "learning_rate": 8.263518223330697e-05,
97
+ "loss": 0.7801,
98
  "step": 30
99
  },
100
  {
101
+ "epoch": 4.615384615384615,
102
+ "eval_loss": 0.7456648945808411,
103
+ "eval_runtime": 1.3352,
104
+ "eval_samples_per_second": 14.979,
105
+ "eval_steps_per_second": 2.247,
106
  "step": 30
107
  },
108
  {
109
+ "epoch": 5.384615384615385,
110
+ "grad_norm": 0.2123967206389714,
111
  "learning_rate": 5.000000000000002e-05,
112
+ "loss": 0.7752,
113
  "step": 35
114
  },
115
  {
116
+ "epoch": 5.384615384615385,
117
+ "eval_loss": 0.7378687262535095,
118
+ "eval_runtime": 1.3301,
119
+ "eval_samples_per_second": 15.037,
120
+ "eval_steps_per_second": 2.256,
121
  "step": 35
122
  },
123
  {
124
+ "epoch": 6.153846153846154,
125
+ "grad_norm": 0.3816302523405475,
126
  "learning_rate": 2.339555568810221e-05,
127
+ "loss": 0.6727,
128
  "step": 40
129
  },
130
  {
131
+ "epoch": 6.153846153846154,
132
+ "eval_loss": 0.738702654838562,
133
+ "eval_runtime": 1.3319,
134
+ "eval_samples_per_second": 15.016,
135
+ "eval_steps_per_second": 2.252,
136
  "step": 40
137
  },
138
  {
139
+ "epoch": 6.923076923076923,
140
+ "grad_norm": 0.20189352065054944,
141
  "learning_rate": 6.030737921409169e-06,
142
+ "loss": 0.7328,
143
  "step": 45
144
  },
145
  {
146
+ "epoch": 6.923076923076923,
147
+ "eval_loss": 0.7381218075752258,
148
+ "eval_runtime": 1.3341,
149
+ "eval_samples_per_second": 14.992,
150
+ "eval_steps_per_second": 2.249,
151
  "step": 45
152
  },
153
  {
154
+ "epoch": 7.6923076923076925,
155
+ "grad_norm": 0.29727669934654843,
156
  "learning_rate": 0.0,
157
+ "loss": 0.6869,
158
  "step": 50
159
  },
160
  {
161
+ "epoch": 7.6923076923076925,
162
+ "eval_loss": 0.7382391095161438,
163
+ "eval_runtime": 1.3314,
164
+ "eval_samples_per_second": 15.022,
165
+ "eval_steps_per_second": 2.253,
166
  "step": 50
167
  },
168
  {
169
+ "epoch": 7.6923076923076925,
170
  "step": 50,
171
+ "total_flos": 1447302955008.0,
172
+ "train_loss": 0.982821364402771,
173
+ "train_runtime": 185.3525,
174
+ "train_samples_per_second": 4.316,
175
+ "train_steps_per_second": 0.27
176
  }
177
  ],
178
  "logging_steps": 5,
 
192
  "attributes": {}
193
  }
194
  },
195
+ "total_flos": 1447302955008.0,
196
  "train_batch_size": 4,
197
  "trial_name": null,
198
  "trial_params": null